Decomposition of a Schur-constant model and its applications (Q1023101): Difference between revisions

An \(n\)-dimensional positive random vector \((\tau_{1}, \dots,\tau_{n})\) is said to be Schur-constant with survival function \(S\) if \(P(\tau_{1}>t_{1}, \dots, \tau_{n}>t_{n})=S(t_{1}+\dots +t_{n})\) for all \((t_{1},\dots, t_{n})\in {\mathbb R}^{n}_{+}\). The authors give a necessary and sufficient condition for a random vector to be Schur-constant without any additional assumptions. Next they discuss the measures of dependence between the components of Schur-constant models, focusing on Pearson's correlation coefficient, Kendall's tau and Spearman's rho. They also consider the connection between multivariate stochastic orders for Schur-constant vectors \((\tau_{1}, \dots,\tau_{n})\) and univariate stochastic orders for \(T_{n}=\sum_{i=1}^{n}{\tau_{i}}\). The theoretical results of the paper are applied to modeling random payments with Schur-constant inter-payment times, and to deriving formulas for some quantities connected with the average payment of a life portfolio with Schur-constant residual life times.